tristan_cordeboeuf
I have to say I'm confused about what you're doing, haha.
From the docs on Stretch Bias:
Set’s the imaginary rest length of the muscle to be (1+stretchBias) times the actual rest length. The line of action will not excite when the curve is this length. Setting stretchBias to a positive number will make the fiber excite even when the curve is not contracted.
The second sentence is probably relevant to what you're seeing.
I've had many discussions internally about the 'muscles contracted at rest' topic. Some of these discussions have taken place only inside my own head. 😀
I'm sure someone will jump in if they have another idea but where I'm currently at is: there's no way to do this with fibers, because a fiber won't help you know what the 'uncontracted' shape should look like. It's only helpful for determining what a contracted shape should look like.
Your best bet is a restshape, as james says.